On Tue, 2 Mar 2004, Max Zaitsev wrote:
> > It's not at all clear what goes wrong with the Genesys controllers. It
> > might be related to the speed of the data transfers; maybe it doesn't like
> > getting too much data too quickly. Oddly, most of the people with
> > problems like yours have found that max_sectors = 128 works.
> >
> I'm mot really familiar with the USB interface nuts and bolts, but it appears
> to me that what we see concerns more the transport layer (if it exists) than
> the usb-storage driver. The Genesis controller does not seem to cope either
> with too large data frames or with too frequent ones. Requesting less data
> while reading sort of fixes the problem by making the Genesis contoller to
> send less data at once. But it looks like the host controller (either on
> hardware or software level) is able to optimise the traffic by a sort of
> caching. So I guess that the host controller glues smaller data frames into
> bigger ones and the Genesis controller is not able to cope with them. Does
> this sound reasonable?
No, the host controllers don't do anything like that. They (and
usb-storage too) just pass requests through as they arrive.
> The thing is that the device is fully functional under Win2k (I have a
> dual-boot setup, so all the hardware is exactly the same). It's a real pain
> for me as a linux fun to have it not running under my main OS :-((
>
> I'm willing to help to track down the problem by providing all sorts of debug
> info. I'm rather fit with C/C++ to do certain things on my own, but never did
> anything on the kernel/device driver level. Just let me know what shell I do.
Well, if you're up for it, here's something that might prove very helpful.
There's a debugging package for USB devices that runs under Windows 2000
called USB Snoopy. You can download it from
<http://www.wingmanteam.com/usbsnoopy/>
It produces voluminous logs (to put it mildly), but seeing exactly how
much data Windows sends, and how quickly it sends it, could tell us what
we need to know. After installing it and enabling the data capture, try
doing the sort of large write that failed under Linux.
Alan Stern
-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
